On Tue, Dec 14, 2021 at 08:16:41AM +0200, Adrian Hunter wrote: > commit c29d9792607e67ed8a3f6e9db0d96836d885a8c5 upstream. > > The space allowed for new attributes can be too small if existing header > information is large. That can happen, for example, if there are very > many CPUs, due to having an event ID per CPU per event being stored in the > header information. > > Fix by adding the existing header.data_offset. Also increase the extra > space allowed to 8KiB and align to a 4KiB boundary for neatness. > > Signed-off-by: Adrian Hunter <adrian.hunter@xxxxxxxxx> > Cc: Jiri Olsa <jolsa@xxxxxxxxxx> > Link: http://lore.kernel.org/lkml/20211125071457.2066863-1-adrian.hunter@xxxxxxxxx > Signed-off-by: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx> > [Adrian: Backport to v5.15] > Signed-off-by: Adrian Hunter <adrian.hunter@xxxxxxxxx> > --- > tools/perf/builtin-inject.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c > index 6ad191e731fc..d454f5a7af93 100644 > --- a/tools/perf/builtin-inject.c > +++ b/tools/perf/builtin-inject.c > @@ -819,7 +819,7 @@ static int __cmd_inject(struct perf_inject *inject) > inject->tool.ordered_events = true; > inject->tool.ordering_requires_timestamps = true; > /* Allow space in the header for new attributes */ > - output_data_offset = 4096; > + output_data_offset = roundup(8192 + session->header.data_offset, 4096); > if (inject->strip) > strip_init(inject); > } > -- > 2.25.1 > Now queued up, thanks. greg k-h